Stefan Vukmanovic
Founder, Starko
Stefan Vukmanovic is the founder of Starko, the software factory that turns customer feedback into merged pull requests. He built the support tools first, the AI helpdesk, ticketing and knowledge base, and now builds the factory on top of it: feedback becomes recommendations, Starko AI builds the accepted ones, and the team reviews and merges the pull requests. He writes about running that loop in production.
Posts by Stefan Vukmanovic
AI Slop in Software: How to Stop Shipping Features Nobody Needs
The worst AI slop in software isn't bad code. It's good code for something your customers never needed. Here's how to keep a team with coding agents tied to what customers actually do.
Conductor Is Joining Meta: What Changes and Where to Move
Meta acquired Conductor on November 4, 2026. Here are the shutdown dates, what to export before November 20, and how to pick a Conductor alternative for your team.
Background Coding Agents Compared (November 2026): Triggers, Sandboxes, Output and Real Pricing
Nine background coding agents compared on where work starts, where it runs, what it hands back and how you pay. Every price checked on November 3, 2026.
Factory AI Droids: What They Do, Pricing and 8 Alternatives
Factory Droids explained: surfaces, Missions, credits and pricing, the Droid CLI, and eight alternatives compared by where work starts and what comes back.
How to Build a Software Factory: 8 Steps From Feedback to Merge
A practical guide to building an AI software factory: what goes in, where agents run, how runs are verified and who merges. With the settings we use in Starko.
Warp Factories Explained: Oz, Pricing, and How It Compares
What Warp Factories is, how it relates to Warp Oz, what it costs, and how it compares with Factory and Starko on where work starts and who writes the spec.
Starko MCP Server: Your Workspace in Claude, Cursor, VS Code
Starko is now an MCP server. Connect Claude, Cursor or VS Code with your own login, pick one workspace, and use the same 75 tools as the Starko chat, with the same approvals.
What Shipped in Starko: 30 October to 1 November 2026
Forty-two pull requests merged in three days, most of them built by Starko AI and every one merged by a person: ⌘K sessions, media in the ticket editor, folder import, faster threads.
Halving Starko Dashboard Load Time: Three Rounds of Fixes
Every read went through a Server Action, which the browser dispatches one at a time. Moving reads to a Route Handler and patching realtime into the cache cut hard loads from 2255 ms to 1097 ms.
Starko AI Sets Up Its Own Sandbox: Node 24, Docker, Secrets
Starko's sandbox now ships Node 24 and Docker Compose, and Starko AI runs a setup session that installs, starts your app, proves it with a screenshot and asks for the secrets only you can give.
Ask, Fix or Prototype: Talking to Starko AI, and the Handoff
Starko AI in a ticket comment now does what you pick: Ask answers and proposes a run, Fix makes the comment the brief, Prototype publishes mockups. A finished run ends with what you need to do.
Page Errors Become Bug Tickets, With No Model in the Path
A JavaScript error on a customer's page becomes one bug ticket per fingerprinted group, filed by a deterministic rule with no model call. How the grouping, noise filters and caps work.
What Is a Software Factory? AI Software Factories Explained
What a software factory is, where the term comes from, what makes the 2026 AI version different, what a dark software factory is, and how to measure one.
From Support Ticket to Pull Request: How a Starko AI Run Works
Everything between accepting a recommendation and merging the pull request: triage, brief, warm sandbox, agent session, independent verification, reviewer agent and the evidence panel.
Why Starko Measures Cost per Merged Pull Request
Seats count people, resolutions count conversations, tokens count effort. None of them tell you whether a coding agent changed your product. Why merged pull requests are the honest unit for judging an agent, and what the factory dashboard shows.
WhatsApp Business API for Customer Support: The Complete 2026 Guide
How the WhatsApp Business API actually works, what it costs, and how to use it for customer support without the setup headaches. From the team building Starko.
Intercom Competitor for Small Business: Is Starko Better?
Comparing Intercom alternatives for a small business? See how Starko compares on pricing, omnichannel support, AI and workflow fit, how it turns customer reports into code fixes, and when Intercom wins.
AI Chatbot Hallucinations: Why Support Bots Get Answers Wrong
Learn what AI chatbot hallucinations are, why customer support bots generate incorrect answers, and how businesses can reduce accuracy risks.
Conversational AI and Revenue Growth: How Much More Can It Sell?
Slow, fragmented communication quietly costs you sales. Here's how conversational AI lifts conversions, cuts churn, and turns customer engagement into revenue.
How to Build a Knowledge Base Your Customers Actually Use
A step-by-step guide to building a knowledge base your customers and your AI actually use, from someone who builds support software for a living: structure, writing, and the analytics loop that keeps it useful.
AI Customer Support Software: A Practical 2026 Guide
What AI customer support software actually does, how the pricing models really work, and how to evaluate it, written by the founder of a platform that runs it in production.
Help Desk & Ticketing Systems: The Complete 2026 Guide
How modern help desk ticketing actually works when AI handles the routine first: the workflow, the metrics that matter, and how to choose one. From the founder of Starko.
Knowledge Base Software: Build a Support Hub That AI Can Actually Use
Why the knowledge base is the real foundation of AI support, how to structure one so it deflects tickets, and how to choose the software. From the founder of Starko.
Shared Inbox Software: Why support@ Breaks and What Replaces It
When a shared mailbox stops working, what shared inbox software actually fixes, and how to choose one that's ready for AI. From the founder of Starko.
Scale Without Hiring: AI Customer Support for Personalized, Sales-Driven Conversations
Responding fast and accurately is the difference between a one-time buyer and a churned customer. AI customer support can compress response times, reduce repetitive work, and route opportunities to sales without asking you to hire a dozen new agents.
How AI Knowledge Bases Shape Support
Learn how AI knowledge bases, omnichannel helpdesk workflows, and support automation help you handle more tickets, faster-without hiring at the same pace.